AutoML Leaderboard

Best model	name	model_type	metric_type	metric_value	train_time
	1_Baseline	Baseline	logloss	2.05573	2.62
	2_DecisionTree	Decision Tree	logloss	7.02917	26.27
	3_Linear	Linear	logloss	2.80413	17
	4_Default_Xgboost	Xgboost	logloss	1.66433	22.96
	5_Default_NeuralNetwork	Neural Network	logloss	2.38751	2.91
	6_Default_RandomForest	Random Forest	logloss	1.8498	27.79
the best	Ensemble	Ensemble	logloss	1.58303	0.37

AutoML Performance

AutoML Performance Boxplot

Features Importance

features importance across models

Spearman Correlation of Models

models spearman correlation

Summary of 4_Default_Xgboost

<< Go back

Extreme Gradient Boosting (Xgboost)

n_jobs: -1
objective: multi:softprob
eta: 0.075
max_depth: 6
min_child_weight: 1
subsample: 1.0
colsample_bytree: 1.0
eval_metric: mlogloss
num_class: 8
explain_level: 2

Validation

validation_type: split
train_ratio: 0.75
shuffle: True
stratify: True

Optimized metric

logloss

Training time

21.7 seconds

Metric details

	0	1	2	3	4	5	6	7	accuracy	macro avg	weighted avg	logloss
precision	1	0	0.285714	0.2	0.416667	0.4	0.444444	0.666667	0.444444	0.426687	0.385626	1.66433
recall	0.8	0	0.333333	0.2	1	0.4	0.8	0.4	0.444444	0.491667	0.444444	1.66433
f1-score	0.888889	0	0.307692	0.2	0.588235	0.4	0.571429	0.5	0.444444	0.432031	0.390865	1.66433
support	5	9	6	5	5	5	5	5	0.444444	45	45	1.66433

Confusion matrix

	Predicted as 0	Predicted as 2	Predicted as 3	Predicted as 4	Predicted as 5	Predicted as 6	Predicted as 7
Labeled as 0	4	0	0	0	0	1	0
Labeled as 1	0	4	1	4	0	0	0
Labeled as 2	0	2	2	1	1	0	0
Labeled as 3	0	0	1	1	2	1	0
Labeled as 4	0	0	0	5	0	0	0
Labeled as 5	0	1	1	1	2	0	0
Labeled as 6	0	0	0	0	0	4	1
Labeled as 7	0	0	0	0	0	3	2

Learning curves

Permutation-based Importance

Confusion Matrix

Normalized Confusion Matrix

ROC Curve

Precision Recall Curve

SHAP Importance

SHAP Dependence plots

Dependence 0 (Fold 1)

SHAP Dependence from fold 1

Dependence 1 (Fold 1)

SHAP Dependence from fold 1

Dependence 2 (Fold 1)

SHAP Dependence from fold 1

Dependence 3 (Fold 1)

SHAP Dependence from fold 1

Dependence 4 (Fold 1)

SHAP Dependence from fold 1

Dependence 5 (Fold 1)

SHAP Dependence from fold 1

Dependence 6 (Fold 1)

SHAP Dependence from fold 1

Dependence 7 (Fold 1)

SHAP Dependence from fold 1

SHAP Decision plots

Worst decisions for selected sample 1 (Fold 1)

SHAP worst decisions from Fold 1

Worst decisions for selected sample 2 (Fold 1)

SHAP worst decisions from Fold 1

Worst decisions for selected sample 3 (Fold 1)

SHAP worst decisions from Fold 1

Worst decisions for selected sample 4 (Fold 1)

SHAP worst decisions from Fold 1

Best decisions for selected sample 1 (Fold 1)

SHAP best decisions from Fold 1

Best decisions for selected sample 2 (Fold 1)

SHAP best decisions from Fold 1

Best decisions for selected sample 3 (Fold 1)

SHAP best decisions from Fold 1

Best decisions for selected sample 4 (Fold 1)

SHAP best decisions from Fold 1

<< Go back

Summary of 3_Linear

<< Go back

Logistic Regression (Linear)

n_jobs: -1
num_class: 8
explain_level: 2

Validation

validation_type: split
train_ratio: 0.75
shuffle: True
stratify: True

Optimized metric

logloss

Training time

15.8 seconds

Metric details

	0	1	2	3	4	5	6	7	accuracy	macro avg	weighted avg	logloss
precision	1	1	0	0.5	0.25	0.5	0.333333	1	0.4	0.572917	0.598148	2.80413
recall	0.8	0.111111	0	0.2	1	0.6	0.4	0.4	0.4	0.438889	0.4	2.80413
f1-score	0.888889	0.2	0	0.285714	0.4	0.545455	0.363636	0.571429	0.4	0.40689	0.379458	2.80413
support	5	9	6	5	5	5	5	5	0.4	45	45	2.80413

Confusion matrix

	Predicted as 0	Predicted as 1	Predicted as 2	Predicted as 3	Predicted as 4	Predicted as 5	Predicted as 6	Predicted as 7
Labeled as 0	4	0	1	0	0	0	0	0
Labeled as 1	0	1	2	0	5	0	1	0
Labeled as 2	0	0	0	1	4	1	0	0
Labeled as 3	0	0	1	1	0	2	1	0
Labeled as 4	0	0	0	0	5	0	0	0
Labeled as 5	0	0	0	0	1	3	1	0
Labeled as 6	0	0	0	0	3	0	2	0
Labeled as 7	0	0	0	0	2	0	1	2

Learning curves

Coefficients

Coefficients learner #1

	0	1	2	3	4	5	6	7
intercept	-1.39388	1.73114	0.506197	0.153146	0.100163	-0.281434	0.0385996	-0.853927
STUDENTID	-1.008	-0.703841	-0.576327	-0.477876	1.12866	1.36647	0.828255	-0.557334
AGE	-0.371017	0.641589	0.19443	0.567653	-0.371999	0.374551	-0.0487999	-0.986407
GENDER	-1.33092	-0.106411	0.192949	-0.272061	-0.234198	-0.428529	0.591353	1.58782
HS_TYPE	-0.369423	0.223793	0.111948	-0.478621	-0.528692	0.0941249	0.195016	0.751854
SCHOLARSHIP	-0.32765	0.513163	0.314177	-0.606248	-0.372437	0.309768	0.192738	-0.0235118
WORK	0.0443235	-0.254999	-0.332436	-0.274234	-1.03072	0.668561	0.622325	0.557175
ACTIVITY	-0.058176	-0.0183316	-0.163904	1.15	-0.587842	-0.384828	-0.264073	0.327158
PARTNER	-0.885389	0.0343999	1.29895	-0.524792	-0.137497	-0.35788	0.278401	0.293804
SALARY	0.424804	-0.00289076	0.0635743	0.295713	0.309141	-0.594782	-0.763054	0.267494
TRANSPORT	0.979139	0.356042	-0.198099	-0.0101227	-0.253968	0.203104	-0.522273	-0.553823
LIVING	0.136181	-0.391756	-0.0397337	-0.00800259	-0.856374	0.160385	0.849858	0.149442
MOTHER_EDU	-0.185124	0.257722	0.329284	-0.725598	0.233806	-0.262943	0.338406	0.014447
FATHER_EDU	0.321277	0.242976	-0.592236	0.327589	-0.353369	-0.0112746	-0.572894	0.637932
#_SIBLINGS	0.433621	0.0597382	-1.07512	0.183132	-0.0112075	0.539391	0.200668	-0.330219
KIDS	0.333439	-0.668519	0.405059	-0.343735	-0.584913	0.262002	-0.0636369	0.660304
MOTHER_JOB	0.447296	-0.440123	-0.0625127	0.754162	0.222982	-0.379387	-0.508108	-0.0343083
FATHER_JOB	-0.0313009	0.194407	0.00193913	0.414445	-0.619556	0.0673695	0.716868	-0.744172
STUDY_HRS	-0.0212173	-0.128914	-0.442671	0.735249	0.292306	-0.157808	-0.215086	-0.0618582
READ_FREQ	0.156082	-0.141015	-0.66132	-0.525773	-0.0387386	-0.317747	0.16582	1.36269
READ_FREQ_SCI	0.272789	-0.583188	-0.00373658	0.585848	-0.208909	0.669762	-0.00638575	-0.72618
ATTEND_DEPT	0.692602	0.0296079	-0.324169	-0.156695	0.0350442	0.0299578	-0.333834	0.0274851
IMPACT	0.136096	-0.115533	0.71144	0.32824	0.241842	-0.542135	-0.404387	-0.355563
ATTEND	0.104396	0.189752	-0.0988823	-0.660452	0.523454	0.310416	0.17183	-0.540513
PREP_STUDY	0.207056	-0.0219448	-0.0230728	0.178665	-0.241055	0.128904	-0.416607	0.188054
PREP_EXAM	-0.190491	0.157574	0.144561	-0.0890285	-0.068846	-0.433839	0.256097	0.22397
NOTES	-0.497763	0.683808	-0.07469	-0.0955776	0.124605	0.327817	-0.529187	0.0609868
LISTENS	-0.354801	0.0608543	-0.231499	0.1876	0.139163	0.111537	-0.0402591	0.127404
LIKES_DISCUSS	-0.0585298	-0.416849	-0.108927	0.495583	0.227761	-0.188257	-0.654527	0.703745
CLASSROOM	0.556064	-0.15675	-0.506561	0.340012	0.119746	-0.415952	0.412695	-0.349253
CUML_GPA	-0.503702	-0.860902	-0.96946	0.030416	0.674193	1.08424	0.80864	-0.263422
EXP_GPA	-0.246761	0.122702	0.00953066	0.995763	-0.478225	0.0855919	-0.560582	0.0719805
COURSE ID	-0.485709	-0.85564	-0.747735	0.187174	0.327883	-0.732382	1.39552	0.910888

Permutation-based Importance

Confusion Matrix

Normalized Confusion Matrix

ROC Curve

Precision Recall Curve

SHAP Importance

SHAP Dependence plots

Dependence 0 (Fold 1)

SHAP Dependence from fold 1

Dependence 1 (Fold 1)

SHAP Dependence from fold 1

Dependence 2 (Fold 1)

SHAP Dependence from fold 1

Dependence 3 (Fold 1)

SHAP Dependence from fold 1

Dependence 4 (Fold 1)

SHAP Dependence from fold 1

Dependence 5 (Fold 1)

SHAP Dependence from fold 1

Dependence 6 (Fold 1)

SHAP Dependence from fold 1

Dependence 7 (Fold 1)

SHAP Dependence from fold 1

SHAP Decision plots

Worst decisions for selected sample 1 (Fold 1)

SHAP worst decisions from Fold 1

Worst decisions for selected sample 2 (Fold 1)

SHAP worst decisions from Fold 1

Worst decisions for selected sample 3 (Fold 1)

SHAP worst decisions from Fold 1

Worst decisions for selected sample 4 (Fold 1)

SHAP worst decisions from Fold 1

Best decisions for selected sample 1 (Fold 1)

SHAP best decisions from Fold 1

Best decisions for selected sample 2 (Fold 1)

SHAP best decisions from Fold 1

Best decisions for selected sample 3 (Fold 1)

SHAP best decisions from Fold 1

Best decisions for selected sample 4 (Fold 1)

SHAP best decisions from Fold 1

<< Go back

Summary of 2_DecisionTree

<< Go back

Decision Tree

n_jobs: -1
criterion: gini
max_depth: 3
num_class: 8
explain_level: 2

Validation

validation_type: split
train_ratio: 0.75
shuffle: True
stratify: True

Optimized metric

logloss

Training time

25.1 seconds

Metric details

	0	1	2	3	4	5	6	7	accuracy	macro avg	weighted avg	logloss
precision	0	0	0	0	0	0	0	0.111111	0.0444444	0.0138889	0.0123457	7.02917
recall	0	0	0	0	0	0	0	0.4	0.0444444	0.05	0.0444444	7.02917
f1-score	0	0	0	0	0	0	0	0.173913	0.0444444	0.0217391	0.0193237	7.02917
support	5	9	6	5	5	5	5	5	0.0444444	45	45	7.02917

Confusion matrix

	Predicted as 0	Predicted as 1	Predicted as 2	Predicted as 4	Predicted as 7
Labeled as 0	0	2	2	1	0
Labeled as 1	0	0	0	5	4
Labeled as 2	0	0	0	5	1
Labeled as 3	0	0	0	0	5
Labeled as 4	0	3	1	0	1
Labeled as 5	0	0	1	2	2
Labeled as 6	1	0	0	1	3
Labeled as 7	0	0	0	3	2

Learning curves

Decision Tree

Tree #1

Rules

if (STUDENTID <= 74.5) and (STUDENTID > 9.5) and (PARTNER > 1.5) then class: 2 (proba: 38.89%) | based on 36 samples

if (STUDENTID <= 74.5) and (STUDENTID > 9.5) and (PARTNER <= 1.5) then class: 1 (proba: 33.33%) | based on 36 samples

if (STUDENTID > 74.5) and (STUDENTID > 87.5) and (FATHER_JOB > 2.5) then class: 4 (proba: 38.89%) | based on 18 samples

if (STUDENTID > 74.5) and (STUDENTID > 87.5) and (FATHER_JOB <= 2.5) then class: 7 (proba: 57.14%) | based on 14 samples

if (STUDENTID > 74.5) and (STUDENTID <= 87.5) and (SALARY <= 1.5) then class: 5 (proba: 80.0%) | based on 10 samples

if (STUDENTID <= 74.5) and (STUDENTID <= 9.5) and (GENDER > 1.5) then class: 7 (proba: 87.5%) | based on 8 samples

if (STUDENTID <= 74.5) and (STUDENTID <= 9.5) and (GENDER <= 1.5) then class: 0 (proba: 50.0%) | based on 8 samples

if (STUDENTID > 74.5) and (STUDENTID <= 87.5) and (SALARY > 1.5) then class: 2 (proba: 40.0%) | based on 5 samples

Permutation-based Importance

Confusion Matrix

Normalized Confusion Matrix

ROC Curve

Precision Recall Curve

SHAP Importance

SHAP Dependence plots

Dependence 0 (Fold 1)

SHAP Dependence from fold 1

Dependence 1 (Fold 1)

SHAP Dependence from fold 1

Dependence 2 (Fold 1)

SHAP Dependence from fold 1

Dependence 3 (Fold 1)

SHAP Dependence from fold 1

Dependence 4 (Fold 1)

SHAP Dependence from fold 1

Dependence 5 (Fold 1)

SHAP Dependence from fold 1

Dependence 6 (Fold 1)

SHAP Dependence from fold 1

Dependence 7 (Fold 1)

SHAP Dependence from fold 1

SHAP Decision plots

Worst decisions for selected sample 1 (Fold 1)

SHAP worst decisions from Fold 1

Worst decisions for selected sample 2 (Fold 1)

SHAP worst decisions from Fold 1

Worst decisions for selected sample 3 (Fold 1)

SHAP worst decisions from Fold 1

Worst decisions for selected sample 4 (Fold 1)

SHAP worst decisions from Fold 1

Best decisions for selected sample 1 (Fold 1)

SHAP best decisions from Fold 1

Best decisions for selected sample 2 (Fold 1)

SHAP best decisions from Fold 1

Best decisions for selected sample 3 (Fold 1)

SHAP best decisions from Fold 1

Best decisions for selected sample 4 (Fold 1)

SHAP best decisions from Fold 1

<< Go back

Summary of 5_Default_NeuralNetwork

<< Go back

Neural Network

n_jobs: -1
dense_1_size: 32
dense_2_size: 16
learning_rate: 0.05
num_class: 8
explain_level: 2

Validation

validation_type: split
train_ratio: 0.75
shuffle: True
stratify: True

Optimized metric

logloss

Training time

1.9 seconds

Metric details

	0	1	2	3	4	5	6	7	accuracy	macro avg	weighted avg	logloss
precision	0.5	0.4	0	0.2	0.5	0.5	0.666667	0.2	0.377778	0.370833	0.365185	2.38751
recall	0.4	0.222222	0	0.2	0.4	0.8	0.8	0.4	0.377778	0.402778	0.377778	2.38751
f1-score	0.444444	0.285714	0	0.2	0.444444	0.615385	0.727273	0.266667	0.377778	0.372991	0.356944	2.38751
support	5	9	6	5	5	5	5	5	0.377778	45	45	2.38751

Confusion matrix

	Predicted as 0	Predicted as 1	Predicted as 2	Predicted as 3	Predicted as 4	Predicted as 5	Predicted as 6	Predicted as 7
Labeled as 0	2	0	0	0	0	0	0	3
Labeled as 1	0	2	1	2	1	2	0	1
Labeled as 2	0	2	0	0	1	1	1	1
Labeled as 3	1	0	1	1	0	0	1	1
Labeled as 4	0	0	0	1	2	0	0	2
Labeled as 5	0	0	1	0	0	4	0	0
Labeled as 6	0	1	0	0	0	0	4	0
Labeled as 7	1	0	0	1	0	1	0	2

Learning curves

Permutation-based Importance

Confusion Matrix

Normalized Confusion Matrix

ROC Curve

Precision Recall Curve

<< Go back

Summary of 1_Baseline

<< Go back

Baseline Classifier (Baseline)

n_jobs: -1
num_class: 8
explain_level: 2

Validation

validation_type: split
train_ratio: 0.75
shuffle: True
stratify: True

Optimized metric

logloss

Training time

1.9 seconds

Metric details

	0	1	2	3	4	5	6	7	accuracy	macro avg	weighted avg	logloss
precision	0	0.2	0	0	0	0	0	0	0.2	0.025	0.04	2.05573
recall	0	1	0	0	0	0	0	0	0.2	0.125	0.2	2.05573
f1-score	0	0.333333	0	0	0	0	0	0	0.2	0.0416667	0.0666667	2.05573
support	5	9	6	5	5	5	5	5	0.2	45	45	2.05573

Confusion matrix

	Predicted as 0	Predicted as 1	Predicted as 2	Predicted as 3	Predicted as 4	Predicted as 5	Predicted as 6	Predicted as 7
Labeled as 0	0	5	0	0	0	0	0	0
Labeled as 1	0	9	0	0	0	0	0	0
Labeled as 2	0	6	0	0	0	0	0	0
Labeled as 3	0	5	0	0	0	0	0	0
Labeled as 4	0	5	0	0	0	0	0	0
Labeled as 5	0	5	0	0	0	0	0	0
Labeled as 6	0	5	0	0	0	0	0	0
Labeled as 7	0	5	0	0	0	0	0	0

Learning curves

Confusion Matrix

Normalized Confusion Matrix

ROC Curve

Precision Recall Curve

<< Go back

Summary of Ensemble

<< Go back

Ensemble structure

Model	Weight
4_Default_Xgboost	3
5_Default_NeuralNetwork	1
6_Default_RandomForest	1

Metric details

	0	1	2	3	4	5	6	7	accuracy	macro avg	weighted avg	logloss
precision	1	1	0.285714	0.5	0.454545	0.6	0.333333	0.666667	0.488889	0.605032	0.633045	1.58303
recall	0.8	0.111111	0.333333	0.2	1	0.6	0.8	0.4	0.488889	0.530556	0.488889	1.58303
f1-score	0.888889	0.2	0.307692	0.285714	0.625	0.6	0.470588	0.5	0.488889	0.484735	0.455491	1.58303
support	5	9	6	5	5	5	5	5	0.488889	45	45	1.58303

Confusion matrix

	Predicted as 0	Predicted as 1	Predicted as 2	Predicted as 3	Predicted as 4	Predicted as 5	Predicted as 6	Predicted as 7
Labeled as 0	4	0	0	0	0	0	1	0
Labeled as 1	0	1	4	1	3	0	0	0
Labeled as 2	0	0	2	0	2	1	1	0
Labeled as 3	0	0	0	1	1	1	2	0
Labeled as 4	0	0	0	0	5	0	0	0
Labeled as 5	0	0	1	0	0	3	1	0
Labeled as 6	0	0	0	0	0	0	4	1
Labeled as 7	0	0	0	0	0	0	3	2

Learning curves

Confusion Matrix

Normalized Confusion Matrix

ROC Curve

Precision Recall Curve

<< Go back

Summary of 6_Default_RandomForest

<< Go back

Random Forest

n_jobs: -1
criterion: gini
max_features: 0.9
min_samples_split: 30
max_depth: 4
eval_metric_name: logloss
num_class: 8
explain_level: 2

Validation

validation_type: split
train_ratio: 0.75
shuffle: True
stratify: True

Optimized metric

logloss

Training time

26.5 seconds

Metric details

	0	1	2	3	4	5	6	7	accuracy	macro avg	weighted avg	logloss
precision	1	0.363636	0	0	0	0	0.16	0.25	0.244444	0.221705	0.229394	1.8498
recall	0.4	0.444444	0	0	0	0	0.8	0.2	0.244444	0.230556	0.244444	1.8498
f1-score	0.571429	0.4	0	0	0	0	0.266667	0.222222	0.244444	0.18254	0.197813	1.8498
support	5	9	6	5	5	5	5	5	0.244444	45	45	1.8498

Confusion matrix

	Predicted as 0	Predicted as 1	Predicted as 2	Predicted as 4	Predicted as 6	Predicted as 7
Labeled as 0	2	1	2	0	0	0
Labeled as 1	0	4	0	0	5	0
Labeled as 2	0	1	0	0	5	0
Labeled as 3	0	0	0	0	4	1
Labeled as 4	0	4	0	0	0	1
Labeled as 5	0	1	0	0	4	0
Labeled as 6	0	0	0	0	4	1
Labeled as 7	0	0	0	1	3	1

Learning curves

Permutation-based Importance

Confusion Matrix

Normalized Confusion Matrix

ROC Curve

Precision Recall Curve

SHAP Importance

SHAP Dependence plots

Dependence 0 (Fold 1)

SHAP Dependence from fold 1

Dependence 1 (Fold 1)

SHAP Dependence from fold 1

Dependence 2 (Fold 1)

SHAP Dependence from fold 1

Dependence 3 (Fold 1)

SHAP Dependence from fold 1

Dependence 4 (Fold 1)

SHAP Dependence from fold 1

Dependence 5 (Fold 1)

SHAP Dependence from fold 1

Dependence 6 (Fold 1)

SHAP Dependence from fold 1

Dependence 7 (Fold 1)

SHAP Dependence from fold 1

SHAP Decision plots

Worst decisions for selected sample 1 (Fold 1)

SHAP worst decisions from Fold 1

Worst decisions for selected sample 2 (Fold 1)

SHAP worst decisions from Fold 1

Worst decisions for selected sample 3 (Fold 1)

SHAP worst decisions from Fold 1

Worst decisions for selected sample 4 (Fold 1)

SHAP worst decisions from Fold 1

Best decisions for selected sample 1 (Fold 1)

SHAP best decisions from Fold 1

Best decisions for selected sample 2 (Fold 1)

SHAP best decisions from Fold 1

Best decisions for selected sample 3 (Fold 1)

SHAP best decisions from Fold 1

Best decisions for selected sample 4 (Fold 1)

SHAP best decisions from Fold 1

<< Go back